For team leads · updated October 2026
Team leads comparing Lindy for a whole team hit the same two questions: who gets billed, and who owns the result. Lindy answers with a shared pool of credits and a seat per person. Kortix is the open-source alternative: one system your team owns and reviews.
# kortix.yamltriggers: - slug: daily-digest type: cron cron: '0 0 9 * * 1-5' prompt: Summarize yesterday's support tickets and open a change request with the digest.Session work reaches main through a change request. Merge is default-deny for agents.
Lindy lists the work a team can push off its plate; on Kortix the same job is one session, and what it produces lands as a change request a person reviews.
One session reads the connected sources, builds the report, and opens it as a change request a person reviews.
| Region | Plan | Actual | Var |
|---|---|---|---|
| Northwind | 810 | 742 | +67 |
| Globex | 180 | 198 | -18 |
| Total | 990 | 940 | +49 |
A session reads the call notes, edits the records, and lands the edits as a change request you read as a diff.
+2 −2
A session sorts the morning mail, drafts the replies, and files the rest, all in a change request a person approves.
A session reads the call notes, writes the scope and the estimate, and opens the draft as a change request for review.
A cron trigger starts a session on the clock, and the brief lands as a change request instead of an email nobody can trace.
Each layer is a file or a command in the repo, and each one is something the team owns rather than rents.
Agents, skills, memory, connectors and triggers are files in one repo, versioned and diffable.
acme-co/ your whole company, as filesagents/ who does the workskills/ how your company does a jobmemory/ what it has learned so farkortix.yaml rules, triggers, connectors, machine3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP. Connector credentials are brokered server-side and never enter the machine.
agents: support-triage: file: agents/support-triage.md connectors: allAny model provider with your own keys, or the ChatGPT plan you already pay for, or sign in with your OpenCode Console account for OpenCode Zen and Go.
model: autoper agent · per session · per messagebring your own key, or use ourssign in with the ChatGPT plan you already pay forOpenCode is the harness. A model on its own answers; the harness gives it planning, tool use and multi-step runs it finishes.
A model on its own answers.The harness gives it planning, tool useand multi-step runs it actually finishes.Powered by OpenCode, configured by a file in the repo.One isolated sandbox per session. Each session has its own isolated machine and branch.
Each session boots its own isolated Linux machinewith your repo and your tools already on it.Session id, sandbox id and branch nameare one and the same string.Session work reaches main through a change request. Merge is default-deny for agents.
project (git repo + kortix.yaml)└─ session ──> cloud computer on a branch └─ the OpenCode agent works └─ change request ──> you review & merge ──> mainOne person tries an AI tool, then three, then the whole team. Two things change at once: the bill, and who can read or undo what the tool produced. Lindy is built for the first person and grows by adding seats. Kortix is built so the whole team shares one system it owns.
Lindy is hosted, and it describes its own product as "your AI teammate that plugs into your tools, learns how you work, and takes real work off your plate." It keeps data in its cloud and says the data is encrypted, never sold and never used to train models. Lindy's pricing page says "everyone draws from one shared pool of credits," credits cost "about a cent each," and admins "can top up at $10 per 1,000." The same page states the seat rule plainly: everyone who signs up, joins the workspace or @mentions Lindy in Slack takes a billed seat. Add a colleague and you add a seat; a single mention draws the same pool down. Monthly credits refresh each cycle and do not roll over, and Lindy pauses when the pool runs low until the credits reset. The arithmetic for a whole team is on what Lindy's pricing means for a team.
Kortix is the open-source AI Operating System: agents, skills, company memory and connectors live in one git repo the team owns. A session runs one agent in its own sandbox on a branch named after the session. The agent commits and opens a change request, and a person reviews the diff and merges it. The report, the updated record and the drafted reply survive as files in the repo, next to the instructions that produced them, so the team reads them, diffs them and rolls any of it back. The full side by side is on the Kortix vs Lindy page.
Lindy keeps the work in its cloud and holds approvals on actions with outside impact, such as sending, updating or posting. Kortix keeps the work in a repo your team reads first, with the tools an agent may touch set to Allow, Ask or Block. When the data has to stay on the team's own network, self-hosting runs the same open-source stack on your own box and is free. The other options for a first team tool are compared on Lindy alternatives.
Short answers about seats, review and where the work lives.
Yes. Kortix has per-resource permissions for people and agents, with roles, groups, and an audit trail.
Approval gates you set, off until you set them. An agent reaches a tool only through the policy on it: Allow, Ask or Block. Session work reaches main through a change request, and merge is default-deny for agents.
Yes. Run it on Kortix Cloud, in your VPC, or on your own on-prem network, and self-host is free. Lindy's pages describe a hosted product with no self-host option and data held in Lindy's cloud.
Any model provider with your own keys, the ChatGPT plan you already pay for, or sign-in with your OpenCode Console account for OpenCode Zen and Go.
In the web app, on desktop and mobile, in the CLI, and in Slack or Microsoft Teams. A message starts a session, and follow-ups stay in it.
Start with one job, one session, one change request.
Open source · Any model, your keys · Self-host, VPC, or on-prem